410 Add identifiable abort ids to Protocol errors

[hridambasu]

Closes #410

[marsella]

Hm, we'll have to figure out how to throw protocol failures correctly from ZK proof validation. Realistically, if a proof fails to verify it's the sender's fault and we already know that (even for proofs in presign). But the proofs don't know who their sender is. We might need to change the way we do error handling when proof validation fails so that we can augment protocol errors with the sender PID (e.g. we might want a separate error enum for "internal cryptography" modules, including paillier, ZKPs, curve points, etc., that we have to convert at the call site to a meaningful external-facing error). cc @gatoWololo.

I actually see a couple of follow-up issues that should come out of this:

• Handle the actual identifiable cases in presign (the output round requires additional work to identify a cheater, but earlier rounds can do it with our current information)
• Redesign errors for ZKP module (or cryptography more broadly) so we can assign blame at the calling level
• Eventually, design + implement identifiable abort for the presign and sign output rounds.
• Add a warning to the README that explains that we do not have identifiable abort fully implemented at this time.

I'll write these up shortly.

[marsella]

I made an epic #486 with some of the issues above in it. So now we have a placeholder for these changes.

Another note on this PR: I think we should add a note to the README to explain the current state of ID abort -- that is, we sometimes report blame when it's easy to assign, but we miss many cases and users should not rely on that field to be complete at this time.

[marsella]

Thanks, changes look great!

There's some extraneous code from another PR that needs to get removed before this can be merged. Please take out those lines.